mols3131d/mols-agent-assets
4 skills4 installsGitHub
Install
npx skills add https://github.com/mols3131d/mols-agent-assetsSkills in this repo
1Agent Asset StudioAgent Asset Studio documents how solo builders and skill authors structure routing skills when agent capability libraries outgrow flat folders. It prescribes domain-prefixed filenames, distinguishes verb skills for single actions from object skills for multi-action domains and place-based hubs that index many sub-skills. The router pattern keeps SKILL.md lightweight: entry logic reads INDEX.csv and hands off to the right workflow file instead of inlining every procedure. That design targets inference cost and context bloat—common pain when dozens of skill directories each carry duplicate scripts. Naming rules also say nested sub-skills can omit redundant domain tokens when the parent already signals the domain. While the excerpt focuses on conventions and hub architecture, the intent is journey-wide: anytime you expand agent tooling, you reuse the same studio/console pattern rather than rescattering capabilities. Pair it with your repo’s actual INDEX and asset manifests when implementing.1installs2Code Builder Colonycode-builder-colony packages architectural-patterns guidance for solo builders who are past the prototype stage and need a repo that screams the business domain instead of looking like a framework dump. The skill applies when you are structuring complete applications, defining service layers, or aligning with screaming and hexagonal ideas—grouping code under domains like user or billing with predictable entry points (handlers), business logic (services), and persistence (repositories). It deliberately excludes one-off scripts, bare configuration, and tiny utilities where extra layers would only add noise. Directory rules keep nesting shallow and names free of tech-polluted suffixes such as user_db. For indie SaaS and API projects, invoking this skill early in Build saves refactor pain during Ship review and Operate iteration because agents and humans share the same layout contract.1installs3Code Reviewer ColonyCode Reviewer Colony packages general review-principles as an agent skill for solo and indie builders who need repeatable feedback on diffs and PRs without ad-hoc taste arguments. It walks reviewers through readability and maintainability (naming, single responsibility, control flow), defensive programming (validation at boundaries, null and index safety, explicit error handling), testability (decoupling from DB and APIs), and performance hygiene such as avoiding redundant work. The skill fits Claude Code, Cursor, Codex, and similar agents when you are shipping features, reviewing agent output, or hardening a small codebase before release. It is methodology, not an automated linter: the agent applies the checklist to the change under review and cites concrete issues. Pair it with your repo’s architecture docs when the skill references colony architectural patterns. Confidence is high for ship/review workflows; it does not replace security scanners or CI.1installs4Document StudioDocument Studio is an agent-oriented documentation kit backed by Python enums and models for ADR, PRD, spec, and kanban DocTypes, plus a rich DocStatus vocabulary and TaskPriority levels. Solo and indie builders use it when they want Claude Code, Cursor, or Codex to emit repo-native planning artifacts with validated YAML frontmatter instead of unstructured chat notes. The Frontmatter class wraps parsed metadata; Document and validate_frontmatter tie content to allowed types and statuses. Helper routines normalize naming, format YAML lists, and only create files when missing so iterative agent runs do not overwrite approved specs. It fits early Validate scoping, ongoing Build PM work, Ship launch checklists captured as docs, and Operate ADR updates when architecture changes—always as structured markdown the rest of the toolchain can grep and link.1installs